sign in sign up
<<
Home , ActBlue, Ann Kirkpatrick, Arizona Senate, Bradley Byrne, Club for Growth, Danny Tarkanian

In the Shadow of Trump: How the 2016 Presidential Contest Affected House and Senate Primaries

Prepared for the 2017 State of the Parties Conference, Akron, Robert G. Boatright, Clark University [email protected]

The presidential race did not quite monopolize all of the uncivil or bizarre moments of the summer of 2016. One of the more interesting exchanges took place in in August of 2016, during the weeks before the state’s Senate primary election.

Senator John McCain, always a somewhat unpredictable politician, has had difficulties in his last two primaries. Perhaps because he was perceived as having strayed too far toward the political center, or perhaps simply because his presidential bid had created some distance between McCain and Arizonans, he faced a vigorous challenge in 2010 from conservative talk show host and former Congressman J. D. Hayworth. McCain ultimately beat back Hayworth’s challenge, 56 percent to 32 percent, but only after a bitter campaign in which McCain spent a total of over $21 million and abandoned much of his “maverick” positioning and presented himself as a staunch conservative and a fierce opponent of (Steinhauer 2010). His task was made easier by his ability to attack Hayworth’s own checkered career in Congress.

In 2016, McCain again faced a competitive primary opponent, physician, Tea Party activist, and two-term State Senator . Ward, like Hayworth, argued that McCain was not conservative enough for Arizona. Ward was (and is), however, a decade younger than Hayworth, and her shorter tenure in political office made it harder for McCain to attack her. And although Ward had several policy disagreements with McCain, she personalized them to a greater degree than Hayworth had. She discussed McCain’s friendly relationship with during the time the two served together in the Senate, and she made little mention of McCain’s ties to and other anti-establishment Republicans.

On July 18, 2016, , in what may well have been an offhand remark to an crowd, referred to McCain’s reluctance to support him. Trump told the crowd that McCain, who had been a prisoner of war in Vietnam, was not a war hero, because war heroes do not get caught. Although many in the media were aghast at this comment, both Ward and McCain’s Democratic opponent, Congresswoman , marveled at McCain’s lack of a response. McCain’s failure to break with Trump after this, said Kirkpatrick, demonstrated that he was not actually much of a maverick (Meyer 2016). Ward, without directly mentioning the Trump comments, announced within the next week that McCain should not serve out the rest of his term; he should step down “as soon as possible.”

Although Trump ultimately endorsed McCain in the primary, Ward seems to have capitalized on the Trump – McCain fight. During the final month of the campaign she sought to frame McCain as a weak candidate – he was too old, she said, and he was unlikely to be a vigorous advocate for Arizona. Ward was the beneficiary of $7.8 million in Super PAC

1 expenditures; among the largest donors to Ward-friendly super PACs was , the hedge fund billionaire who was the largest donor to pro-Trump groups. The two major super PACs aiding her, her own KelliPAC and the Courageous Conservatives PAC, aired several advertisements describing McCain as being “weak” on immigration and crime. Despite McCain’s mostly pro-gun voting history, he also was the subject of attack ads from .

Ward probably took the “weakness” argument a bit too far; in an August 25 interview on CNN, she told reporter Chuck Todd that “John McCain has fallen down on the job. He's gotten weak. He's gotten old. . . I want to give him the best birthday present ever: the gift of retirement.” Ward went on to say that, statistically, McCain was unlikely to serve out his term, and he would not be able to serve effectively. When challenged by Todd on this, she said that “I'm a physician. I see the physiological changes that happen in normal patients again and again and again over the last 20, 25 years, so I do know what happens to the body and the mind at the end of life” (Watkins 2016).

As future events would show, Ward may well have had a point about McCain’s health, but the blunt way in which she framed the issue likely harmed her more than it helped. In an election full of overheated rhetoric, Ward was able to draw some attention to herself. Would Ward have run this sort of a campaign had Donald Trump not been in the race? Would she have done better? Worse? We cannot, of course, answer these questions, but it is worth speculating about such things. On the one hand, McCain’s interaction with Trump may well have made him more vulnerable, even if Trump’s comments were unfair. Trump, after all, had exchanges with and in the 2016 debates where, despite Trump’s boorish behavior, he seems to have successfully wounded his opponents. Perhaps Trump wounded McCain and Ward was the beneficiary.

On the other hand, there are always some political candidates who say bad, or unfair, or ridiculous things about their opponents. Nobody expected Ward to win, and she didn’t. By many measures she was a weaker candidate than Hayworth had been. She may well have run exactly the same campaign regardless of who the Republican presidential nominee was. Maybe she was just a desperate challenger, saying what it took to get attention, or maybe she was tactless and would have been tactless no matter what. Her public statements after the election were consistent with what she had said in the campaign. She announced during the days after Trump’s July 2017 brain cancer diagnosis that the only reasonable approach for Arizona Governor to take would be to select her for McCain’s Senate seat once he died (Resnik 2017). She also announced her 2018 to Arizona’s other Senator, , using language similar to what she had used against McCain; on August 10, 2017, she told “I, as well as others, are frustrated with Senator Flake’s weak style. He hasn’t accomplished a thing in D.C., and his values do not align with the people of Arizona” (Goodman 2017).

This odd little moment in the primary illustrates some of the challenges one faces in trying to determine what the 2016 presidential race meant for other candidates on the ballot. The drama of this election was so inescapable during the summer that it is hard to imagine it did not have an effect on other primary elections. Yet determining its effects is

2 difficult. Were some of the candidates acting Trump-like because Trump had given them the idea, or is this just who they were?

In this paper, I explore some of the ways that particular types of candidates might have been advantaged or disadvantaged by the presidential race. There are some slight effects. But the largest effect was a negative one – there is some evidence that the 2016 congressional primaries were largely ignored because of the presidential race. This is important because congressional primaries in 2010 through 2014 were not ignored, and the turmoil of these years’ primaries has had major effects on the Republican Party in particular. The decline in the salience of primaries may have consequences going forward.

Measuring Primaries and Primary Effects

In order to understand the effects of the Trump campaign on the 2016 primaries, a few details on what is “typical” in House and Senate primaries are in order, as are some details on the timing of House and Senate primaries. To make a long story short, primaries usually are quiet affairs. Downballot primaries traditionally exhibit low turnout; since 2002, turnout in congressional primaries has average approximately twenty percent of the electorate. Few of those who do vote tend to know very much about House and Senate primary candidates at all (Boatright 2014, ch. 3; Gerber, Huber, Biggers, and Hendry 2016). Since 1994, there has only been one election cycle (2010) in which more than ten percent of incumbents faced a primary opponent who held them to less than 75 percent of the vote. In a typical election, there are no more than three or four incumbents who lose their primaries, and when incumbents do lose, it is generally because of a scandal or some other highly publicized bit of malpractice. A select number of open seat primaries exhibit higher turnout; challenger primaries (that is, those which select a candidate who will take on the incumbent in the general election) tend to have even lower turnout and lower levels of competition, in large part because gerrymandering has reduced general election competitiveness, and few general election challengers have a chance of winning.

In recent years, there has certainly been more attention paid by the national political media to primaries. This is in part due to aggressive efforts by interest groups to back ideological or anti-establishment candidates in the primaries. As I have documented (Boatright 2013a), this sort of strategy has created the perception that moderates of both parties risk “getting primaried” if they engage in bipartisan policymaking or adopt positions at odds with the preferences of the party base. The establishment of super PACs has, as exemplified by the Kelli Ward challenge described above, enabled idiosyncratic wealthy donors to almost singlehandedly put candidates whom they dislike on the defensive.

These efforts may attract attention, but they remain . That is, stories about primaries tend to be more influential in shaping public views than have actual events. The Republican Party, in particular, has seen a spate of particularly divisive House and Senate primaries over the past decade, while Democratic primaries have been less competitive. Despite the defeat of House Majority Leader , however, there were signs that this Republican unrest had

3 begun to subside in 2014. Republican primaries remained more contentious than Democratic primaries, and the number of Republican House and Senate incumbents facing opponents continued to climb, but, Cantor’s upset aside, most “establishment” incumbents comfortably beat back their opponents. This was, I have argued, to be expected – primary election tumult is positively correlated with the level of general election interparty competition, and there was no reason to expect the “wave” elections of 2006-2012 to be followed by still more general election instability. The 2014 election dealt a blow to the story line that primary tumult was threatening the Republican Party.

Going into 2016, then, one would have expected that Republican primaries would be more competitive than Democratic primaries, but that the level of chaos in Republican primaries might be expected to decline in comparison to 2012 and 2010. Such a development (a development which, as I will show here, did largely come to pass) would represent an important change in American politics. Many analyses of political polarization in the have pointed to primary elections as an important barometer of the movement of both parties, and the Republican Party in particular, away from the political center. The Democratic gains of 2006 and 2008 occurred in tandem with unrest among liberals; this unrest resulted not only in general election victories but in challenges to incumbent Democratic centrists such as Senator Joseph Lieberman. Republican victories in 2010 were propelled in part by the same Tea Party activists who spearheaded primary challenges to incumbent Republicans such as Alaska Senator Lisa Murkowski and Senator . These years were also characterized by upsets in a number of open seat and challenger primaries, such as ophthalmologist and political neophyte1 ’s defeat of the establishment favorite, Attorney General Trey Grayson in the 2010 Republican Senatorial primary and ’s defeat of the more moderate (or at least more “establishment-preferred”) candidate, Lieutenant Governor David Dewhurst in the 2012 Republican primary.

Collectively, these campaigns have created the perception that both parties’ establishments are on the defensive. This is a perception that organized interests have carefully worked to foster; as I discuss in Getting Primaried, ideological interest groups have concentrated their resources in a small number of primaries in order to draw attention to themselves. A decline in primary election unrest might, then, have reduced the perceived incentives for legislators to polarize, and might have reduced the clout that ideological groups – the Tea Party, in particular – play in elections.

Such predictions, however, do not take into account any sort of spillover effects from the presidential campaign. In one sense, Trump is of a piece with the sorts of Republican insurgents who have run against incumbents in GOP primaries. He cast his campaign as a run against the party establishment and against career politicians. Unlike these candidates, however, Trump did not cast his campaign in strictly ideological terms; he ran as an outsider, but not necessarily as a Tea Party candidate or a candidate with commitments to the tenets of conservative or far right organizations. It was hypothesized at various points during the primaries that Trump led his own sort of movement. More specifically, it was claimed that Trump

1 Paul is the son of former congressman and presidential candidate , so he certainly knows about politics. Rand Paul had never held or sought political office, however. 4

a) had highly enthusiastic supporters whose views diverged from those of traditional Republicans; it was unclear what these voters (if they indeed existed) would do in other primaries (see Railey 2016; Phillips 2016); b) would upend conventional understandings about the relationship between campaign fundraising and victory; because Trump raised so much less than his opponents, he would inspire other poorly funded candidates to adopt similar rhetorical techniques or efforts to accrue free media coverage (Everett 2016); and c) would encourage other Trump-like candidates (what New York Times columnist Frank Bruni (2016) dubbed “mini-Trumps”) to run for office.

We cannot prove whether or not any of these things did take place, but we can look at Republican congressional primary competition in 2016, as compared to previous years, to see whether Trump’s presence on the ballot seems correlated with any unusual patterns in candidate emergence, the level of competitiveness, or spending in the primaries. We are able to do this both by comparing all 2016 Republican primaries to those of previous years, by comparing them to Democratic primaries, and by comparing concurrent primaries – those in states where the presidential and state primaries were held on the same day – to nonconcurrent ones. This also gives us a means of comparing early and late primaries, in that concurrent primaries tend to take place earlier than do nonconcurrent ones. In short, we can explore whether the 2016 House and Senate primaries were unusual, and speculate about whether any abnormalities had anything to do with the unusual presidential race. These questions loosely map on to three variables that are measurable across time: turnout (which says something about voter enthusiasm), competition (which would be higher if more people sought to run for office), and spending. This paper proceeds by looking at how the 2016 looked with regard to these variables, and it then turns to the task of measuring some more complicated matters such as candidate rhetoric and how candidates responded to themes and events of the presidential campaign.

This is an important task for four reasons. First, political scientists have an incomplete understanding of the relationship between presidential and nonpresidential primaries. While we know that voter turnout is higher when congressional primaries are held in conjunction with presidential ones, we do not necessarily know what this means for the types of candidates who are elected or for the characteristics of congressional campaigns. In 2008, there was speculation about an “Obama effect” on candidates who shared a primary ballot with him (Boatright 2013, 36-37). The presence of an unconventional candidate such as Donald Trump can give us some insight into this – did he help or hurt Republicans (and Democrats) who were also on the ballot? And how did candidates adjust their strategies based on expectations of what might happen?

House Primary Competition in 2016

Data on voting and election results cannot conclusively address the hypotheses above, but they can get us close to ascertaining whether anything unusual happened in downballot primaries.

5

Turnout

The major question regarding the 2016 electorate is, again, whether voters in 2016 were any different than they would have been had the presidential candidates been different. There are several ways of going about answering this question: looking at levels of turnout; looking at the characteristics of the voters; and looking at characteristics of the people who voted for particular presidential candidates. There is, regrettably, no nationwide survey data available for 2016 that includes vote choice in House or Senate primaries.2 We must, therefore, make some inferences about what actually took place in congressional primaries.

Let us first consider aggregate primary election turnout in 2016. In any given election year, elections will be held for some but not all Senate seats and governorships, and the specific states where these primaries will be held will not necessarily be representative of the country. Thus, we shall consider turnout in House elections here. Although in some states House primaries are not held when there are fewer than two candidates, considering average turnout percentages for the districts that do hold primaries can give us a way of presenting a national comparison across time. We should not assume that voters in House primaries showed up to vote because there was a House primary (in fact, they likely showed up for other reasons). But if we are interested in the effect of the presidential primaries, it is appropriate to compare House primaries that were and were not held on the same day as the presidential primary. Figures 1 and 2 provide such comparisons.

[Figures 1 and 2 about here]

Figure 1 shows voter turnout in presidential year concurrent and nonconcurrent Republican primaries. Data here are drawn from the Boatright/Moscardelli dataset on congressional primary voting.3 Turnout is always higher in presidential primaries than it is in nonpresidential ones. Turnout in the 2016 Republican primaries was higher than it had been in any previous year, which meant that turnout in concurrent House primaries was also higher than it had been in any prior year.4 Roughly 50 percent more voters took part in the 2016 Republican presidential primaries than took part in the 2012 presidential primaries, and roughly 35 percent more voters took part in the concurrent downballot primaries. The gap between concurrent and nonconcurrent primaries was also wider than had been the case in any election year for which we have good data. The people who showed up to vote for president, in other words, were either not traditional Republican primary voters or were particularly uninterested in Republican downballot candidates in 2016. The former seems a bit more probable. The figure shows turnout for spring primaries on the right; if we assume that presidential primary competition disappears by summer,

2 The 2010 and 2014 CCES both included questions about congressional primary vote choice, but the 2012 and 2015 surveys did not. The ’s Primary Project (https://www.brookings.edu/series/the-primaries- project/) drew in part on exit polls from select congressional districts, but the data from the Brookings study is not yet available, and most publicly available data related to this project concerned candidates, not voters. 3 Available at https://wordpress.clarku.edu/primarytiming/. 4 Estimates for presidential primary voters are drawn from the Pew factsheet on primary turnout, at http://www.pewresearch.org/fact-tank/2016/06/10/turnout-was-high-in-the-2016-primary-season-but-just-short-of- 2008-record/. 6 a measure for the spring captures more elections in which voters actually have a meaningful choice.

Before we assume that this shows some sort of great influx into the Republican Party, however – and before we go assuming that it is an influx of Democrats – consider Figure 2. Turnout in the Democratic primaries was higher than in Republican primaries; it was higher than it had been since 1992; and the gap between concurrent and nonconcurrent Democratic primaries was also larger. So it is not evident that the Trump voters were renegade Democrats, and it is also not evident that the Republican turnout numbers (which after all amounted to less than 12 percent of the electorate) showed some remarkable level of enthusiasm. Both parties’ core voters were more excited than they had been for quite some time, but this still amounts to a small share of the electorate.

Just because more people voted in 2016, however, we should not necessarily assume that the 2016 voters brought markedly different views or characteristics to bear on their voting choices. It is difficult to make direct comparisons between the 2012 and 2016 electorates, but Table 1 provides some comparisons. Note that two different surveys are used here; for 2012, I use the Pew survey of Republican primary voters; for 2016 I use the Cooperative Congressional Election Study (CCES). Pew has not published a survey of 2016 voters, and the 2012 CCES common content does not include questions on primary voting.

[Table 1 about here]

Despite differences in question wording, sampling, and sample size, it seems evident that the 2016 primary electorate was not necessarily different in ways that would obviously influence voting for congress. If the comparisons here are accurate, the 2016 Republican electorate was more male, but also younger and less white. It could be that women crossed party lines to vote for Hillary Clinton (or men crossed party lines to vote for a Republican). But these characteristics would not seem to have an obvious connection to support for different types of Republican candidates. In fact, the changes in the characteristics of Republican voters are striking because minorities and younger voters voted overwhelmingly for Clinton in the general election (although not to the degree that they had in 2008 or 2012). If the 2016 voters were different in terms of ideology, this might be of consequence. Unfortunately, the differences in the categories offered and the question wording make it hard to draw precise comparisons, but it does not appear that the 2016 voters were either more or less conservative than the 2012 Republican primary voters.

It is perhaps unilluminating to point out that the supporters of different presidential candidates have different views. Data gathered by the Blair Center showed that Trump supporters differed from supporters of other candidates in several important ways, including scoring higher on the center’s “modern sexism” index than all of the other candidates’ supporters except for Ted Cruz’s.5 Other surveys show a wide range of other differences between supporters of the various candidates in terms of geography, age, and other characteristics and beliefs. Such differences are to be expected and occur in any contested primary. What is most consequential for our purposes, however, is how the presidential candidates’ supporters

5 For a summary of findings see https://blaircenter.uark.edu/the-impact-of-modern-sexism/. 7 influenced other elections. If Trump or Cruz supporters voted in past Republican primaries, then their views were not newly inserted into the primary process.

The closest we can get to figuring out how Trump supporters influenced the Republican congressional primaries is to determine whether they are any more or less happy with congress, and with their member of congress, than were the supporters of other candidates. To do this, let us consider the following data from the CCES. The CCES presented respondents with a variety of different support measures and ideology scales. Table 2 includes only responses from citizens who resided in a district represented in the House by a Republican. These voters were asked whether they approved of the way Congress is doing its job, and of the way their representative is doing his or her job.6 Responses are scaled such that a score of 1 would indicate that the respondent strong approved, and a score of 4 would indicate that the respondent strongly disapproved. High mean scores, then, show lower average levels of support. We would assume that a Republican who answered that he or she disapproved on either question would be more likely to vote against the incumbent than would one who approved. As the table shows, however, Trump supporters look just like other Republicans; in fact, it is the Cruz voters who stand out. Cruz voters like Congress less, but like their own representative more, than the supporters of other Republican candidates.

[Table 2 about here]

What this table shows in part, however, is that everybody hates Congress. Clinton supporters in Republican districts are actually more likely to say that they approve of the job their representative is doing than are Republican voters; 24.2 percent of her primary voters approved, or strongly approved, of the job their representative was doing, as compared to 22.0 percent for Trump voters, 21.6 percent for Cruz voters, and 16.7 percent for Kasich voters. Only Marco Rubio’s supporters (27.3 percent) liked their representative better.

Table 3 presents ideological scales of how supporters of the various presidential candidates who reside in Republican congressional districts rate their representatives and the Republican Party more generally. A response here of 1 corresponds to a “very liberal” measure, a 4 corresponds to “moderate” and 7 correspond to “very conservative.” Here again, it is the Cruz voters, not the Trump voters, who stand out. Cruz voters think that the Republican Party as a whole is significantly more moderate than do supporters of any other candidate. Trump voters look just like Rubio and Kasich voters on this measure. Given that Cruz, and Cruz supporters, believed themselves to be more conservative than other candidates and their supporters, it is the Cruz voters who would seem most likely to support a conservative congressional primary opponent. Cruz voters are no different than Trump voters in how they view their own representatives’ ideologies, though, which may suggest that Cruz voters already reside in districts that have Republican representatives to their liking.

[Table 3 about here]

6 Similar questions were asked in some 2012 surveys (including the CCES) but since public support for Congress is not something that would remain consistent across elections (and since Democrats controlled the Senate in 2012) we have not compared responses to this question across elections. 8

It is possible, of course, that Trump voters might have been supportive of Republican candidates who deviated from the party norm in ways other than ideology. What these data show, however, is that there is no systematic way to draw conclusions about this. Absent a district-by-district analysis of congressional primary voting – an analysis which we cannot obtain – there is no way here to conclude that Trump supporters influenced downballot Republican elections. If there was any sort of influx of new voters brought into the primary by Trump (a claim that is also not supported by the data), these voters did not have any inclination to punish Republicans. What seems more likely is that new Republican voters were drawn to the election simply by the excitement and the unsettled nature of the primary; that these voters were not even all Trump supporters; and that these voters considered downballot Republican candidates the same way traditional Republican primary voters did.

Competition

Figure 3 shows patterns over time of primary competition for Democrats and Republicans (shown separately), with the lines in each graph representing incumbent primaries, open seat primaries, and challenger primaries.7 The election year is on the x-axis. The variable on the y- axis is the mean fractionalization index. The fractionalization index is a measure of competitiveness which ranges form 0-1, where higher values are indicative of greater competition or fractionalization of the party primary vote.8 The mean of this is simply the average level of fractionalization across all primaries of a particular type. As this figure shows, there was little change in House primary competitiveness between 2014 and 2016. Republican incumbent primaries became slightly more competitive than 2014, but otherwise there was little change. It is worth noting, in addition, that the increase in primary competitiveness was not necessarily driven by successful “insurgent” candidates. Among the most successful primary challengers in 2014 were mainstream Republicans running with the support of the U.S. Chamber of Commerce and other business groups to reclaim seats that had been won by Tea Party candidates in 2010 or 2012.

[Figure 3 about here]

Candidates’ decisions about whether to run for office are, of course, made well in advance of the presidential race, and there was little reason to expect anyone who decided in mid-2015 or earlier to run for Congress to have gamed out the consequences of running in a year when Donald Trump would be the Republican nominee. It is possible, however, that the

7 That is, primaries that include an incumbent office-holder; primaries for a general election race in which there will be no incumbent, and primaries whose nominee expects to face the incumbent in the general election. 8 Specifically, the fractionalization index is operationalized as 2 2 2 2 F = 1 - ∑ [(C1) + (C2) + (C3) + (C4) . . .] where F is the fractionalization index, C1 is the percentage of the total vote received by the first candidate, C2 is the percentage of the total vote received by the second candidate, and so on. This yields an index where a one candidate race has a fractionalization index of zero and a race where two candidates split the vote would have a fractionalization index of 0.5 (or 1 – (0.52 + 0.52)). The larger the number of similarly competitive candidates, the closer the index is to 1 – that is, a race with ten candidates who received ten percent of the vote each would have an index of 1 – [(.1)2 * 10], or 0.9. This index was originally developed by Canon (1978), and is used by, among others, Herrnson and Gimpel (1995) and Hogan (2003). 9 presence of Trump on the ballot on the same day as one’s own primary would influence downballot voting. That is, people inspired to vote for Trump might not be the sorts of people who would reflexively vote for Republican incumbents in downballot races. Figure 4 compares concurrent downballot primaries (primaries held on the same day as the presidential primary) and nonconcurrent primaries, again using the fractionalization index and comparing 2016 to prior election years – although in this case, we are only considering prior presidential election years. While it is important to note that the characteristics of the states holding concurrent or nonconcurrent primaries might differ, the graphs here provide absolutely no evidence that characteristics of the voters who turned out for presidential election days had any meaningful effect on competition in Republican House primaries. If anything, it was the Democrats who had more contentious primaries on these days, despite the overall lack of competitiveness in Democratic primaries.

[Figure 4 about here]

Thus, even if we cannot definitively say anything about whether Trump-like candidates were inspired to run, we can conclude that the Trump voters had no discernible effect on primary competition. The 2016 primaries were, in comparison to the past decade of congressional primaries, rather sedate.

Spending

Next, let us consider spending in the 2016 House primaries. We can understand spending in two different ways. First, expectations of victory in the general election may inspire more spending in the primary election. Although we have seen that there was little evidence of an increase in primary competition, candidate spending may have served to increase or decrease the level of competition, or particular types of candidates may have spent money in anticipation of the threat of competition. Second, I noted above that during previous election cycles independent expenditures have become consequential in Republican primaries – and particularly, in incumbent primaries – as mainstream and conservative groups have sparred over the future of the party’s congressional delegation.

Figure 5 shows two different measures of Republican primary spending. The figure on the left distinguishes between incumbent, challenger, and open seat primary spending. In order to exclude primary spending that is most likely aimed at the general election, I limit the cases included in the graph here to primaries that were at least somewhat competitive (using only primaries where the ultimate vote share ratio was 75-25, or its equivalent using the fractionalization measure).9 In these races, open seat primary spending increased, although open seat race spending can vary idiosyncratically, according to the number of open seats in some

9 I use this as a minimal threshold for competitiveness, following my work in Boatright 2013. 25 percent tends to correspond to the amount candidates who file with the FEC receive. Since we cannot conclusively determine whether a prospective nominee is spending money to combat his or her primary or general election foes, there is no way to be certain whether the right amount of money is being included here, but in my 2013 book on primaries I provide a lengthy justification for this with reference to several particular primary elections from the past few election cycles. 10 elections. The other two types of spending, however, declined. The drop in incumbent primary spending is particularly noteworthy, suggesting that even though the fractionalization of Republican incumbent primaries has continued to climb, these races had become less costly. This is a measure of aggregate spending, by incumbent and their primary challengers, suggesting not only that fewer primary challengers spent large sums of money, but that incumbents spent less to ward off such challenges.

[Figure 5 about here]

The right-hand graph in Figure 5 compares spending in concurrent and nonconcurrent primaries, again limiting consideration to somewhat competitive races. Averages for nonconcurrent primaries are far more variable than averages for concurrent ones, in part because the number of somewhat competitive nonconcurrent primaries has fluctuated wildly. There was an average of 14.5 somewhat competitive nonconcurrent primaries in each year from 1984 through 2012, ensuring that one or two very expensive primaries could skew the results, as happened in 2004. There were a total of 66 somewhat competitive nonconcurrent primaries in 2016, however. As the earlier figures on fractionalization show, this larger number did not translate into a higher average level of competition overall, but this figure clearly does indicate that Republican spending in concurrent primaries was unremarkable. Republican incumbents, in other words, do not seem to have worried about their opponents any more than usual.

In sum, the competition and the candidate spending data do not suggest that the 2016 Republican House primaries were more tumultuous than Republican primaries had been in the previous years. If anything, they were less exciting. It seems plausible that the presidential campaign drained whatever excitement might have existed in the Republican downballot races.

As a final piece of suggestive evidence about this, let us briefly consider independent expenditure (IE) data for 2016. IEs, which I discuss in more detail elsewhere (Boatright 2017), generally consist of advertising or other spending which specifically advocates for the election or defeat of a candidate for office. I noted above that the major recipients of support from outside groups in 2016 tended to be either incumbents who had reason to expect primary opposition, or, in a small number of instances, centrist, business-friendly candidates facing Tea Party Republican incumbents. This was the pattern in 2014 as well.

Ten House primary candidates were the beneficiaries of over $500,000 in independent spending, and 38 were the beneficiaries of over $100,000 in independent spending. In many cases, the independent expenditures were made by individual funders with personal ties to the candidate; this is the case for the top Republican recipient of IE support, Republican Amie Hoeber, who was the beneficiary of $1.66 million in expenditures made by a super PAC created and funded by her husband. Centrist Republicans such as (R-OH) and John Shimkus (R-IL) benefitted from large independent expenditures made by business-friendly groups; both of these candidates did face conservative opponents, but these conservative opponents raised far less. Shimkus’s opponent did benefit from slightly over $100,000 in Super PAC spending, mostly from the .

11

Organizations that had been supportive of Republicans outside of the mainstream in the past were present in 2016, but they spent far less and they spent their money either in open seat races or in support of incumbents who they liked but who did not necessarily need very much help. The Club for Growth spent $1.2 million in House primaries (excluding an Ohio special election primary) and $700,000 in Senate primaries. Apart from one open seat primary in , however, the bulk of this money was spent in uncompetitive primaries, and thus was actually money spent to influence the general election. Freedomworks, the wealthiest independent spending group associated with the Tea Party, followed a similar strategy; and the Senate Conservatives Fund, a Super PAC created by former Senator Jim DeMint, spent only $300,000. Each of these groups had spent over – and in some cases well over – $2 million in the 2012 and 2014 primaries.

These (admittedly anecdotal) group examples suggest that it is not just that there was less money in the 2016 primaries; the money came from different places, and organizations that had in the past sought to sow a bit of discord in Republican primaries were restrained in 2016.

Senate Primary Competition in 2016

House primaries provide a useful gauge of general unrest within a party’s primary electorate because the number of them is sufficiently high that the effect of any unusual races does not unduly distort the averages presented here. Averages for Senate primaries, on the other hand, may be influenced by one or two particularly idiosyncratic primaries. However, on average the 2016 Senate primaries were less competitive than they had been since 2008. To take one example, Figure 6 shows average primary vote share for Republican and Democratic incumbents, with a dot for each incumbent (since most incumbents face no opposition at all, the dots at the top of graph represent multiple candidates). There was only one Republican incumbent in any danger of losing his primary – this was Arizona Senator John McCain – and there were two other Republican incumbents who were held to less than seventy percent of the primary vote yet did not have opponents who raised enough money to pose a serious threat. The Senate primaries, in short, also show little evidence of any sort of ideological turbulence.

[Figure 6 about here]

The story is much the same regarding independent expenditures during the primaries, as Figure 7 shows. Despite the smaller number of competitive Democratic primaries, independent spending was substantially higher among Democrats in 2016 than it ever had been. In Senate races, this was largely driven by two competitive open seat primaries in the safe Democratic states of and Maryland. Independent expenditures in Republican primaries, however, fell precipitously in the House and in the Senate. While the decline in the Senate may have been a function of the smaller number of competitive races, the fact that House competition did not drop dramatically but independent spending did suggests a lack of interest in congressional elections on the part of the major Republican-allied and conservative Super PACs. Super PACs are a new enough phenomenon that we do not yet have a sense of what is “normal” Super PAC

12 spending behavior in a presidential year. Yet the independent spending data provide more evidence that Republican donors and groups were not particularly interested in the 2016 primaries.

[Figure 7 about here]

In the Senate, there was only one primary challenger who received more than token support from outside groups; this was Kelli Ward, discussed above. There were few competitive Republican open seat Senate races; some money was spent in the Florida Republican primary but this money ultimately disappeared after the incumbent Senator Marco Rubio abandoned his presidential bid and reentered the race. Primaries in Indiana and attracted multiple strong candidates but independent spending generally supported the candidate with closer ties to the Republican establishment. In the Senate, then, it is hard to argue that the presidential race increased primary activity by nonparty groups.

It is not, then, just that things didn’t happen. As I argued in Getting Primaried and in other work on primaries, the success of the narrative about the threat Republican moderates face has been built on the regular convergence of outside groups in a small, carefully selection number primary elections – elections chosen to send a message to other Republicans. Challenges to Republican moderates (R-PA, 2004 and 2010), Lincoln Chaffee (R- RI, 2006), Lisa Murkowski (R-AK, 2010), Robert Bennett (R-UT, 2010), Richard Lugar (R-IN, 2012) drew news coverage even though few of them were successful. These challenges suggested that conservative groups had the potential to make reelection complicated for any legislator who crossed them. In 2014, four Republicans (, R-TN; , R-MS; Mitch McConnell R-KY , R-KS) faced conservative opponents, but all survived. The effort put in by the Republican Party and organizations like the U.S. Chamber of Commerce created a sort of counternarrative that the business community would protect mainstream Republicans. We thus have two election cycles in a row where the only Republican incumbents who had serious worries about keeping their seats were Republicans who were too conservative. At some point, the notion that there is some sort of threat out there on the political right loses its credibility. 2016 may have represented a step in that direction.

Candidate Characteristics and Strategies in the Shadow of Trump

The primary data, then, suggest that the 2016 downballot primaries were, if anything, rather dull. There was no ideological or anti-establishment wave going on in the Republican Party. Is this an effect, or a lack of an effect? There is ample reason to contend that the presidential race so overshadowed some congressional races that some potentially troublesome Republican candidates were starved of attention. In 2016, any threats to sitting Republicans were destined to be seen not as singular lessons about collateral damage from the presidential campaign. This had the effect of framing any unusual primaries as examples of candidates who had characteristics that were of relevance to the Republican (and to a lesser extent, the Democratic) presidential contest.

13

Tea Party Effects?

Early in the primary season, there was speculation that Trump voters, and perhaps Cruz voters as well, would complicate the renomination efforts of people like Senator or Alabama House members and . All three of these incumbents had been the targets of conservatives in the past, and each did face an opponent who might have served as a rallying point for conservative media figures or interest groups. Yet neither of the two House challengers raised over $200,000, and although Shelby’s Senate challenger raised more ($766,000) he was not financially competitive with Shelby. Most consequentially, there was no independent spending on behalf of any of these candidates.

As the 2016 primaries wore on, the media identified a few candidates who appeared to be vaguely “Trump-like” – candidates such as conservative talk show host Jason Lewis, who ran in a southwestern Minnesota open seat House race, and Darryl Glenn and Robert Blaha, who both ran in the primary. These were candidates with a history of incendiary statements, although each of these candidates had run before, and each had been associated in some way with the . Only Blaha actually supported Trump in the primary; he wound up running a distant third. As his hold on the nomination tightened, Donald Trump also publicly associated himself with two different primary candidates, , an incumbent House member from North Carolina, and , who ran against Speaker of the House . Neither candidate won; Ellmers received just 23.6 percent of the vote in losing an unusual incumbent-vs.-incumbent race brought about by the state’s court-ordered redistricting, and Nehlen received only 15.9 percent of the vote in his challenge to Ryan. Nonparty groups spent $57,000 on behalf of Nehlen; Ellmers was the beneficiary of over $300,000 in independent expenditures, but over $1.1 million was spent in support of her opponent. Spending in the Ellmers race would certainly have taken place without Trump’s involvement in the race. Trump also publicly feuded with Senator John McCain, although he ultimately did support McCain and Ryan. Slightly over $870,000 in independent expenditures were made against McCain, as compared to over $4 million spent in support of McCain.

The track record of candidates who seemed to those in the media to resemble Trump, or who received support from Trump in some fashion, was thus not impressive. Of these candidates, only Lewis won, and it is by no means clear that any resemblance between the two, apart from their shared partisan affiliation, had anything to do with his victory. Lewis was treated by election handicappers as a likely loser for much of his campaign, and he downplayed any efforts to link him to Trump, tersely saying he would support Trump if he were the nominee, on the grounds that he did not want to “turn the over to the Clintons” (Zurowski 2016). In addition, much of the spending in these races may well have been there without Trump’s intervention.

This is not to say, however, that the sorts of political elites who supported Donald Trump were not active in the 2016 primaries. As we saw in the Arizona example above, donors with a history of antagonism toward the Republican status quo, such as the Mercer family, waded into the 2016 primaries. These donors, however, would arguably have behaved in the same way

14 regardless of who the Republican nominee was. Robert Mercer gave over $22 million to Republican candidates in 2016, and he spent on behalf of several Republican primary candidates. But few of these candidates had any ties to Trump, and Mercer had given generously to a wide range of Republicans in 2012 and 2014 as well. The same can be said of the Ricketts family and other Republican Super PAC donors. These donors took a chance on Trump, but they did not seek to create a Trump-esque Republican Party.

It seems that if there was any sort of Trump effect in the primaries, then, it had to do with a stifling of ideological competition within the Republican Party. Almost all of the primaries that attracted spending from ideological interest groups either were in districts that had been redrawn (as was the case in Florida and North Carolina) or were competitive open seat races with multiple strong candidates, as was the case in states such as and Arizona. There were only two House races that featured high profile battles between competing Republican groups: the three-way open seat primary race in New York’s 22nd Congressional District, won by , and the successful challenge to Representative in the 1st District. Tenney was a second-time candidate who had come within six percentage points of defeating moderate Republican Richard Hanna in 2014, and Huelskamp, a prominent Tea Party candidate in 2010, had repeatedly tangled with agricultural and business interests in his district. Both of these races, in other words, featured candidates who had already established a track record of garnering outside support. No new House primary candidates emerged in 2016 with significant backing from outside groups.

One could argue that Trump had little to do with this – that the 2016 primaries were merely a placid set of races. But candidacy decisions are made long before the dynamics of the presidential race become clear, and the candidates were there in 2016. There was the potential for competition. The fact that little money showed up to support this competition was not something candidates could have known in advance, and is likely a decision of independent expenditure groups not to focus on primaries. This could have merely been an acknowledgment that the money would be wasted on elections that would not receive public attention, as perhaps might have been the case in the early, concurrent primaries, or a decision in the later, mostly nonconcurrent primaries to save money for a general election in which Republicans would be on the defensive.

Democratic Primary Candidates and the Trump (and Clinton and Sanders) Effect

And finally, what of the Democrats? In many of the seats expected to be competitive, Democratic donors and interest groups coalesced around promising candidates and effectively winnowed the field before the primary was underway. Table 4 shows this phenomenon: here we see primary election fundraising and independent expenditures in four open seat primaries. Each of the seats listed here was expected to be close, but Democratic money was far more concentrated than was Republican money. The difference between Democrats and Republicans here is not unique to 2016, and it is also not necessarily indicative of the general election results – Republicans won all of these races save for the Nevada one. Over the past decade Democrats have consistently had less competitive primaries than Republicans, and when they have had competitive primaries, they have been among ambitious office-seekers running in safe

15

Democratic states or districts. 2016 was, then, not an aberration for the party, nor is there reason to expect that the nature of the impending general election or the Democratic or Republican presidential primary races had an obvious effect on this dynamic.

[Table 4 about here]

While it was a challenge to find candidates who we can conclusively identify as “mini- Trumps,” the list of candidates associated with is easier to come by. Sanders supporters created an official organization, , to support progressive candidates. Our Revolution has continued to be active since the 2016 election and uses its webpage to steer visitors to the fundraising pages of its endorsed candidates. Daily Kos, the most widely read progressive website, also published a list in April 2016 of 43 House and Senate candidates who had endorsed Sanders.10 Although Sanders did not respond by endorsing all of these candidates during the primary, or at any other point, many did enthusiastically connect themselves to his campaign.

This list is intriguing in its sheer variety. Most of these candidates lost overwhelmingly, but it does not appear that they lost because of their connections to the Sanders campaign. Many were long shot primary or general election candidates, and many were running in heavily Republican districts. ActBlue and other left-leaning netroots websites sought to raise money for them, and a few – most notably, unsuccessful New York House candidate Zephyr Teachout and successful -area open seat candidate Pramila Jayapal11 – were able to establish a national network of small donors. The idea behind the effort to promote candidates linked to Sanders was not necessarily to win elections, but to recruit a stable of candidates who might run again in 2018. The “Our Revolution” organization, headed by liberal luminaries such as former Texas politician Jim Hightower and the leaders of various liberal cause groups, made it clear that their mission was to train candidates for the future. The Sanders campaign was thus presented as an example of a movement that was to persist with our without Sanders’ direct involvement.

There has since the election been talk of an enduring split within the Democratic Party between Sanders supporters and mainstream Democrats; this conflict drew much attention in the competitive race between former Labor Secretary and Minnesota Congressman Keith Ellison for the chairmanship of the Democratic National Committee. Yet there was no evidence that hostilities between the two sides influenced Democratic congressional primaries or general election candidacies in 2016. Given the relative placidity of Democratic primaries, as compared to Republican ones, there is no reason to expect that the Sanders campaign will have an enduring effect in dividing the party or hampering general election campaigns. The surge in announced Democratic candidates for 2018 (Malbin 2017) may indicate that efforts such as Our Revolution are paying off in candidate recruitment, but it is hard to separate the effects of this effort from a surge in candidacies in response to the Trump victory or the perception that the 2018 midterm election appears likely to be a good one for Democratic House candidates.

10 https://www.dailykos.com/stories/2016/4/23/1519069/-Help-the-43-Democrats-running-for-US-House-Senate- seats-that-have-endorsed-Bernie-Sanders. 11 Jayapal, running in a top-two race in the heavily Democratic Seventh district of , did not have a Republican opponent in the general election. 16

Conclusions

In this paper I have sought to document the relatively placidity of the 2016 congressional primaries. The lack of competition, both electoral and financial, suggests that the presidential race choked off the sort of ideological competition that has characterized primaries, and in particular, Republican primaries, for the past decade. This is an important development. Will such competition reemerge in the next few election cycles? It is certainly possible that at least in the next midterm elections congressional primary candidates will have more of a megaphone for their campaigns. On the other hand, the groups that have instigated ideological competition have done so in part to draw attention to themselves and their agendas; given Republican dominance of government, it would seem more complicated for conservative anti-government groups to continue to inspire a national following by railing against moderate Republicans. The phenomenon of “primarying” may well have begun to go into remission.

The data in this paper suggest that there may have been a “Trump Effect” in 2016, but that this effect was not what pundits might have predicted. There was no upside for candidates who linked themselves to Trump, either directly, or in terms of adopting his sort of style of campaigning. If there was a Trump effect, it was manifested in distracting the public from the ideological competition that had characterized Republican elections for nearly a decade.

It is important to remember, however, that candidates make their decisions about whether to run well before the primary season begins. Republicans did not enter the race because of Trump. Some Republicans will no doubt base their decisions about whether to run in 2018 on Donald Trump’s actions in office, but this will of course cut both ways. Some potential Republican candidates will be deterred from running because of concern about the political environment in 2018, and some of the Republican incumbents who have pondered how to respond to Trump’s actions will likely have to think about the implications of that for their reelection bids. The two Republican Senators who appear to be in the most general election jeopardy in 2018 – Arizona Senator Jeff Flake and Nevada Senator – had declared primary opponents by August 2017. McCain’s foe Kelli Ward announced her campaign against Flake, and Danny Tarkanian, a frequent Republican candidate, announced his intention to challenge Heller. Both of these Republican incumbents have been critical of Trump, and Trump has been critical of them. Trump donors have contributed to these challenges, although Trump has not been directly involved (yet).12 Tarkanian, in announcing his bid, stated that “We’re never going to unless we have senators in office that fully support President Trump and his America-First agenda” (Beaumont 2017). Trump himself, if his treatment of Paul Ryan and John McCain in 2016 is any indication, may well avoid directly campaigning against these candidates. Trump chose in the 2017 special election to replace Alabama Senator to ally himself with the most conventional candidate in the race,

12 On August 17, 2017, Trump appears to have endorsed Ward in a tweet, but it is too early to tell whether he will actively support her (see Nakamura and O’Keefe 2017). 17

Luther Strange, over two more incendiary primary foes (Blinder and Martin 2017).13 Neither Trump nor his supporters, then, are taking steps to form any sort of organized effort to reorient the Republican Party.

The legacy of the 2016 primaries, then, will likely be inextricably tied to the fate of the Trump presidency. The Republican Party had a primary problem before 2016; the specific problem it had – challenges from the right, factional conflict between business groups and Tea Party conservatives, a growing unwillingness on the part of the right wing to work with the rest of the Republican Party – may have been crystallized by the Trump candidacy and the difficulties Congress had in legislating in 2017. But in the years to come it will be hard to determine whether Republican conflict has anything to do with these disagreements or whether it is rooted in factors specific to the ideas and political style of Donald Trump. The Democrats, on the other hand, showed some signs of the emergence of similar factional conflict, but Democrats, despite their weakened status, had succeeded in minimizing the effects of these conflicts on intraparty battles, and the prospects for success in 2018 may well continue to help the party paper over these conflicts.

13 Strange, appointed to fill Sessions’ seat after Sessions became Attorney General, finished second in Alabama’s special primary runoff on August 15, 2017. 18

Bibliography

Beaumont, Thomas. 2017. “Nevada Senator Faces Pro-Trump Challenge in GOP Primary.” Washington Post, August 8. Blinder, Alan, and Jonathan Martin. 2017. “Despite Trump’s Support, , Alabama Senator, Struggles to a Primary Finish.” New York Times, August 13. Boatright, Robert G. 2013a. Getting Primaried: The Changing Politics of Congressional Primary Challenges. Ann Arbor, MI: University of Press. Boatright, Robert G. 2013b. “The Voice of American Business: The U.S. Chamber of Commerce and the 2010 Elections.” In Interest Groups Unleashed, ed. Paul S. Herrnson, Christopher J. Deering, and Clyde Wilcox. Thousand Oaks, CA: Press, pp. 31-56. Boatright, Robert G. 2014. Congressional Primary Elections. New York: Routledge. Boatright, Robert G. 2016. “U. S. Interest Groups in a Deregulated Campaign Finance System.” In The Deregulatory Moment?, ed. Robert G. Boatright. Ann Arbor: University of Michigan Press, pp. 71-106. Boatright, Robert G. 2017. “Damage Control: The Trump Effect on Senate General Election Campaigns.” In The 2016 Presidential Election: The Causes and Consequences of a Political Earthquake, ed. Amnon Cavari, Kenneth Mayer, and Richard Powell. Lanham, MD: Lexington Books, pp. 59-76. Bruni, Frank. 2016. “The Misery of the Mini-Trumps.” New York Times, August 28, p. SR3. Canon, Bradley. 1978. “Factionalism in the South: A Test of Theory and a Revisitation of V. O. Key.” American Journal of Political Science 22 (3): 833-848. Chait, Jonathan. 2016. “How do Republican Protect Themselves from Trump?” New York Magazine, May 4. Everett, Burgess. 2016. “Nuisance Challengers Irk GOP Senators.” , April 12. Gerber, Alan S., Gregory A. Huber, Daniel R. Biggers, and David J. Hendry. 2016. “Why Don't People Vote in U.S. Primary Elections? Assessing Theoretical Explanations for Reduced Participation.” Electoral Studies 45 (1): 119–129. Goodman, Ted. 2017. “Jeff Flake’s Primary Challenger to Hire Top Trump Operatives.” The Daily Caller, August 10. On line, http://dailycaller.com/2017/08/10/exclusive-jeff- flakes-primary-challenger-to-hire-top-trump-operatives/. Hermson, Paul S., and James G. Gimpel. 1995. “District Conditions and Primary Divisiveness in Elections.” Political Research Quarterly 48 (1): 101-16. Hogan, Robert E. 2003. “Competition in State Legislative Primary Elections.” Legislative Studies Quarterly 28 (1): 103-126. Malbin, Michael J. 2017. “Does the Opening Predict a Wave?” Brookings Institution Fixgov , July 24. On line, https://www.brookings.edu/blog/fixgov/2017/07/24/does-the- opening-predict-a-wave/. Meyer, Theodoric. 2016. “McCain Battles Trump’s Shadow in Arizona.” Politico, July 19. Nakamura, David, and Ed O’Keefe. 2017. “Deepening GOP Split, Trump attacks GOP Senators Graham, Flake, as ‘Publicity Seeking,’ ‘Toxic.’” Washington Post, August 17. On line, https://www.washingtonpost.com/news/post-politics/wp/2017/08/17/deepening-gop-split- trump-attacks-republican-senators-graham-flake-as-publicity-seeking-toxic/?hpid=hp_hp- more-top-stories_flake-840am%3Ahomepage%2Fstory&utm_term=.cf3ad5b6cd42.

19

Phillips, Amber. 2016. “In Alabama Today, the Trump Effect Threatens a Republican Longtime Senator.” Washington Post, March 1. Railey, Kimberly. 2016. “The First House Primary to Watch in 2016.” , January 21. Resnik, Brahm, 2017. “Former McCain foe urges cancer-stricken senator to quit so she can get seat.” NBC 12 News Arizona, July 21. On line, http://www.12news.com/news/politics/former-gop-foe-urges-cancer-stricken-mccain-to- quit-so-she-can-get-seat/458565749. Steinhauer, Jennifer. 2010. “From Right of Radio Dial, a Challenge to McCain.” New York Times, February 8. Watkins, Eli. 2016. “McCain’s GOP Challenger says he’s too old for the job.” CNN Politics, August 26. On line, http://www.cnn.com/2016/08/25/politics/kelli-ward-john- mccain/index.html. Zurowski, Cory. 2016. “Why Jason Lewis Refuses to Abandon the Sinking Ship that is Donald Trump.” City Pages, October 14.

20

Figure 1: Turnout in Concurrent and Nonconcurrent Republican Primaries, 1980-2016

All Primaries Spring Primaries Only

21

Figure 2: Turnout in Concurrent and Nonconcurrent Democratic Primaries, 1980-2016

All Democratic Primaries Spring Primaries Only

22

Figure 3: Fractionalization by Primary Type and Party, 1970-2016

Republicans Democrats

23

Figure 4: Fractionalization in Concurrent and Nonconcurrent Primaries, by Party, 1972-2016

Republicans Democrats

24

Figure 5: Spending in Republican Presidential Year House Primaries, 1980-2016

25

Figure 6: Incumbent Vote Share in Senate Primaries, by Party, 1978-2016

Republicans

Democrats

26

Figure 7: Independent Expenditures in Primary Elections, 2006-2016

Republicans Democrats

27

Table 1: Comparing the 2012 and 2016 Republican Primary Voters

2012 2016 White 91 87 Over 50 73 64 Under 30 4 8 Male 46 55 4 yr. College Degree or More 35 38 Moderate 21 19 Conservative* 55 53 Very Conservative 21 21

Source: For 2012, Pew Republican Voters Survey, http://www.people- press.org/2016/01/28/2012-republican-primary-voters-more-conservative-than-gop-general- election-voters/; for 2016, CCES. * The Pew survey does not give a “somewhat conservative” option, the CCES does. The CCES total here combines “somewhat conservative” and “conservative” responses. The Pew survey also does not provide “don’t know” responses for ideology questions while the CCES does.

28

Table 2: Selected Performance Evaluations by Presidential Primary Vote Choice

Primary Vote Choice Congress Job Republican N Performance (mean) Incumbent Performance (mean) Trump 3.26 2.67 4,498 Cruz 3.23 2.43 2,404 Kasich 3.36 2.86 942 Rubio 3.08 2.56 919 Another Republican 3.31 2.72 602 Clinton 3.32 3.54 4,801 Sanders 3.60 3.76 3,745 Another Democrat 3.23 3.08 104 Total N 18,015

Mean scores taken from a 1-4 scale where 1 = “strongly approve” and 4 = “strongly disapprove.” “Don’t Know” responses and voters who chose a candidate of another party excluded.

Table 3: Selected Ideology Scores by Presidential Primary Vote Choice

Primary Vote Choice Republican Party Republican N Ideology (mean) Incumbent Seeking Reelection (mean) Trump 5.22 5.20 4,202 Cruz 4.69 5.38 2,343 Kasich 5.78 5.49 892 Rubio 5.54 5.60 894 Another Republican 5.19 5.41 578 Clinton 5.84 5.54 4,412 Sanders 6.23 5.96 3.,520 Another Democrat 5.05 5.30 94 Total N 1,523

Mean scores taken from a 1-7 scale where 1 = “very liberal” and 4 = “very conservative.” “Don’t Know” responses and voters who chose a candidate of another party excluded.

29

Table 4: Primary Competition and Financing, Selected 2016 House Open Seat Races

Candidate Party Vote Receipts Independent Primary General Share Expenditures Result Result MN-2 D 100 $2,524,279 W L Jason Lewis R 48.9 369,431 W W Darlene Miller R 30.7 393,971 John Howe R 13.7 746,801* NV-3 Jacklyn Rosen D 62.00 579,142 W W Jesse Sbaih D 12.80 772,672* Daniel Tarkanian R 32.00 968,356* W L R 24.00 977,367 $2,522,056 R 18.20 378,781 Andrew Matthews R 14.10 254,586 NY-22 Kim Myers D 100 690,679* W L Claudia Tenney R 41.50 257,041 14,975 W W Steven Wells R 33.7 808,241 1,229,862 George Phillips R 24.8 243,982 3,445 MI-1 Lonnie Johnson D 71.60 1,285,143 W L Jerome Cannon D 28.40 48,555 John Bergman R 38.60 309,082* W W Tom Casperson R 32.00 384,189 177,000 Jason Allen R 29.40 309,647 222,682 * Total includes $100,000 or more in self-financing. Democratic candidates in italics. IE totals include money spent in favor of candidate and in opposition to principal opponent.

30